Statistical test and error bars?

Hi! I'm completely new to statistics so I'm a bit stuck! I'v been treating 4 groups of human cells (all originally from the same batch) with different doses of a chemical. I was then looking to see how many of the cells were normal after treatment (had 46 chromosomes) and how many were abnormal (had less or more than 46 chromosomes)in each treatment group.So lists were made of how many chromosomes were contained in each analysed cell. I thought the best way to do this was to do a graph showing the percentage of cells within each group that had 46 chromosomes - and there does appear to be a trend between the 4 groups but I have no idea where to go from here! How would I do error bars on my graph? I calculated the standard deviation for each group from the raw data, do I just use these SDs when I specify the error values? (I wasn't sure if it would be different given I've done my graph as percentages from the original data?). Also are there any statistical tests I could use? I thought maybe ANOVA? But again does it matter that I've only included the data about cells with 46 chromosomes on my graph as a percentage of the total data?
I hope i'm making sense?! And I apologize if these are stupid questions, I'm just getting so confused! Many thanks for any advice :)


Omega Contributor
Correct me if I read your post incorrectly, but it seems like you have 4 treatment groups with n=? cells in each group and your outcome is binary (46 Yes or No).